How do I turn my Claude Artifact into an Android or iPhone app?

You’ve built something with Claude and someone has said, “You should make it an app.” Before doing that, I’d ask whether it actually needs to be one.

A website already works on phones. An app adds store fees, a review process and a second codebase to maintain. It’s worth it when you need something a website can’t do well, such as reliable push notifications or deeper access to the phone’s hardware, or when your customers expect to find you in the app stores. If neither applies, I’d start with the website and see whether anyone asks for more.

Three routes onto a phone

1. Installable website (PWA)

A progressive web app is your normal website with a bit of extra setup, so people can add it to their home screen and open it like an app. There’s no store, no fee and no review process, and every update to the site reaches everyone straight away. For most projects this is where I’d start. What you give up is a listing in the app stores, and a lot of people simply don’t know that you can install a website.

2. Wrapped web app (Capacitor)

Capacitor puts your existing web code inside a thin native shell, which means you can submit it to the stores and use some of the phone’s features. If you already have a working web app, it’s the quickest way into the stores. The thing to be careful of is wrapping a website and expecting it to be approved. Apple can reject apps that don’t do anything more than the website already does.

3. Native app (React Native with Expo)

Here the app is built for the phone from the start. React Native uses the same language and a similar component model to a React artifact, and Expo takes care of most of the build and publishing work. It gives the best experience and the most access to the phone. The cost is that it’s a separate project from your website, so there’s more to build and more to maintain.

Claude can help you work out which of the three fits what you’ve built.

The app still needs a back end

This is the part that tends to get missed. The app is only the screens. If it saves anything, logs anyone in or sends an email, it still needs a server and a database behind it, exactly as a web app does. What your app actually needs goes through how to work out which pieces you need.

What the app stores want

The 14-day closed test on Google Play is easy to miss if you haven’t published before, so line up your testers early and put the time in the plan. Fees and rules change, so check Google Play Console and Apple Developer before you commit to a launch date.

Building on Windows or Linux

Android apps can be built on Windows or Linux without any trouble. iPhone apps are harder, because Apple’s build tools only run on a Mac. If you don’t have one, a cloud service such as Expo EAS Build can do the build for you, although you’ll still need the paid Apple developer account. To start an Expo project:

npx create-expo-app@latest my-app
cd my-app
npx expo start

npx create-expo-app@latest my-app
cd my-app
npx expo start

That prints a QR code. Install Expo Go on your phone, scan the code and the app opens on your phone. If you haven’t installed Node.js yet, the website guide covers it, along with keeping your code in a private GitHub repository, which applies here too.

The order I’d do it in

1. Get the web version live and let real people use it.
2. Find out whether anyone is actually asking for an app.
3. If they are, make the website installable first, as a PWA.
4. If you still need the stores, ask Claude to build the Capacitor or Expo version, using the same back end.
5. Open the store accounts, find your testers and test on real Android and iPhone devices.
6. Publish your privacy policy, write your store listings and submit.

Before you submit, go through the launch checks. All of it applies to apps as well.

If you’re not sure whether you need an app, ask me. If a website will do the job, I’ll say so. See my services, or get in touch for a free 20-minute call.